This is the program for the 2010 Joint Statistical Meetings in Vancouver, British Columbia.

Abstract Details

Activity Number: 33
Type: Contributed
Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
Sponsor: Section on Bayesian Statistical Science
Abstract - #308062
Title: Bayesian Dynamic Item Response Models
Author(s): Xiaojing Wang*+ and James Berger and Donald Burdick
Companies: Duke University and Duke University and MetaMetrics, Inc.
Address: 214 Old Chem, Department of Statistical Science, Durham, NC, 27708,
Keywords: Item response models ; State space models ; Latent variable ; Gibbs sampling ; Forward filtering and backward sampling ; Monotone constraints
Abstract:

Item response models enjoy popularity with large-scale educational and measurement testing and many studies have contributed to its development and applications in static time. However, when there are repeated observations for individuals at different time, a dynamic structure should be built for the latent trait of persons' ability. So this paper proposes a new class of state space models for dichotomous response data, which not only shows serial dependence in the data, but also allows the growth effects of persons' ability to vary in different time lapse. Moreover, a novel Markov Chain Monte Carlo algorithm is suggested for statistical inference, which incorporates isotonic constraints on the latent trait without extra burden for computation. The related simulations and applications prove that our models provide an appealing tool to measure and monitor persons' ability growth.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2010 program




2010 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.